pythia-autopilot
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a safety-first workflow for autonomous operations, emphasizing 'least privilege' by requiring the agent to propose the narrowest possible scope for its tasks.
- [SAFE]: It implements strict boundaries through 'The Iron Law,' which prevents the agent from stretching lease limits or accessing production/protected connections defined in local configuration files.
- [SAFE]: The instructions include a comprehensive 'Red Flags' section that explicitly warns the agent against deceptive behaviors, such as skipping impact assessments or attempting to enable autopilot mode without developer intervention.
- [SAFE]: No malicious code, obfuscation, data exfiltration, or unauthorized privilege escalation patterns were detected.
Audit Metadata